The Role of Hexokinase and Hexose Transporters in Preferential Use of Glucose Over Fructose and Downstream Metabolic Pathways in the Yeast Yarrowia Lipolytica

International Journal of Molecular Sciences

Article The Role of Hexokinase and Hexose Transporters in Preferential Use of Glucose over Fructose and Downstream Metabolic Pathways in the Yeast Yarrowia lipolytica

Piotr Hapeta 1 , Patrycja Szczepa ´nska 1, Tadeusz Witkowski 1, Jean-Marc Nicaud 2 , Anne-Marie Crutz-Le Coq 2 and Zbigniew Lazar 1,*

1 Department of Biotechnology and Food Microbiology, Wrocław University of Environmental and Life Sciences, Chełmo´nskiegoStreet 37, 51-630 Wrocław, Poland; [email protected] (P.H.); [email protected] (P.S.); [email protected] (T.W.) 2 Université Paris-Saclay, INRAE, AgroParisTech, Micalis Institute, 78352 Jouy-en-Josas, France; [email protected] (J.-M.N.); [email protected] (A.-M.C.-L.C.) * Correspondence: [email protected]

Abstract: The development of efficient bioprocesses requires inexpensive and renewable substrates. Molasses, a by-product of the sugar industry, contains mostly sucrose, a disaccharide composed of glucose and fructose, both easily absorbed by microorganisms. Yarrowia lipolytica, a platform for the production of various chemicals, can be engineered for sucrose utilization by heterologous invertase expression, yet the problem of preferential use of glucose over fructose remains, as fructose consumption begins only after glucose depletion what significantly extends the bioprocesses. We Citation: Hapeta, P.; Szczepańska,P.; investigated the role of hexose transporters and hexokinase (native and fructophilic) in this prefer- Witkowski, T.; Nicaud, J.-M.; ence. Analysis of growth profiles and kinetics of monosaccharide utilization has proven that the Crutz-Le Coq, A.-M.; Lazar, Z. The glucose preference in Y. lipolytica depends primarily on the affinity of native hexokinase for glucose. Role of Hexokinase and Hexose Interestingly, combined overexpression of either hexokinase with hexose transporters significantly Transporters in Preferential Use of accelerated citric acid biosynthesis and enhanced pentose phosphate pathway leading to secretion Glucose over Fructose and of polyols (31.5 g/L vs. no polyols in the control strain). So far, polyol biosynthesis was efficient in Downstream Metabolic Pathways in glycerol-containing media. Moreover, overexpression of fructophilic hexokinase in combination with the Yeast Yarrowia lipolytica. Int. J. Mol. hexose transporters not only shortened this process to 48 h (84 h for the medium with glycerol) but Sci. 2021, 22, 9282. https://doi.org/ 10.3390/ijms22179282 also allowed to obtain 23% more polyols (40 g/L) compared to the glycerol medium (32.5 g/L).

Academic Editor: Ismail Fliss Keywords: Yarrowia lipolytica; hexokinase; hexose transporters; glucose; fructose; erythritol

Received: 25 July 2021 Accepted: 25 August 2021 Published: 27 August 2021 1. Introduction The oleaginous yeast Yarrowia lipolytica is an industrial micro-organism used for Publisher’s Note: MDPI stays neutral production of complex biological compounds such as tailor-made lipids [1,2], biofuel pre- with regard to jurisdictional claims in cursors [3] or industrial and therapeutic enzymes [4,5], as well as more compact molecules published maps and institutional afﬁl- such as acidifying agent citric acid [6] and low-calorie sweeteners, including erythritol and iations. mannitol [7,8]. Substrates can be chosen among hydrophobic carbon sources (n-alkanes, oils and fats) and hydrophilic ones, such as glycerol, monosaccharides and organic acids. From the latter, glycerol is a superior carbon source for Y. lipolytica, due to its rapid and preferred use over glucose when supplied in a mixture [9–11]. To date, production of Copyright: © 2021 by the authors. erythritol by this yeast relies on the use of glycerol in the culture medium as it acts as a Licensee MDPI, Basel, Switzerland. stress factor, increasing osmotic pressure required for erythritol overproduction [12]. In This article is an open access article turn, the narrow spectrum of monosaccharides used by wild-type Y. lipolytica strains, which distributed under the terms and includes glucose, fructose and mannose [13], can be extended by genetic engineering to conditions of the Creative Commons galactose and xylose [14–16]. Glucose and fructose are readily available substrates present Attribution (CC BY) license (https:// in plant biomass and industrial by-products such as molasses [17]. Many yeast species, creativecommons.org/licenses/by/ including Saccharomyces cerevisiae, show a preferential uptake of glucose over fructose in 4.0/).

Int. J. Mol. Sci. 2021, 22, 9282. https://doi.org/10.3390/ijms22179282 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2021, 22, 9282 2 of 20

mixed sugar media [18]. This trait is particularly prominent in Y. lipolytica, for which sev- eral experiments enlightened that fructose uptake begins only after glucose is completely or mostly depleted from the medium [19,20]; the molecular mechanism behind remains unknown. In order to enter the metabolism, monosaccharides need first to be taken up by the cell, a step which requires transporters, then to be activated through phosphorylation by hexose kinases. Y. lipolytica possesses six bona fide hexose transporters (encoded by the genes YHT1 to YHT6) belonging to the 24-protein Sugar Porter (SP) family [20]. Among them, Yht1, Yht3 and Yht4 are efficient broad-range hexose transporters enabling glucose, fructose, mannose and galactose uptake as determined by the heterologous complementation study [20]. Yht1 and Yht4 were identified as the main active hexose transporters in this yeast as deletion of both genes prevented growth on the four hexoses [20]. Furthermore, out of the six YHT genes, only YHT1 (YALI0C06424g) and YHT4 (YALI0E23287g) were consistently transcribed during the growth of W29 and H222 strains in glucose- and fructose-based media [20]. These were also the only Yht transporters expressed in media containing glucose, glycerol and a mixture of both under nitrogen-limiting conditions, showing a differential expression pattern depending on carbon source and nitrogen concentration [9]. YHT3 (YALI0F19184g) appeared to be poorly transcribed in the conditions tested. Moreover, the strain-dependent polymorphism in this gene had an impact on the efficiency of the transporter as the gene in Y. lipolytica H222 encoded a much more efficient transporter than its counterpart from the W29 strain [20]. Two enzymes with hexokinase activity have been identified in Y. lipolytica. One of them is a glucokinase, with glucose as the only substrate; the other one phosphorylates glucose, fructose and mannose [21] and is encoded by the YlHXK1 (YALI0B22308g) gene. The primary structure of YlHxk1 has quaint characteristics absent in other known hexokinases. In addition to conserved domains allowing for the action on hexoses, YlHxk1 contains an atypical 37-amino acid loop spanning between glucose- and ATP-binding domains [21], which indirectly influences enzyme activity [22]. The overexpression of YlHXK1 enhanced fructose utilization in the Y. lipolytica strains showing tedious growth on this monosaccharide, such as the wild strain W29, and more generally improved hexose utilization, increased carbon flux through glycolysis, changed the preferential use of glycerol to glucose when both substrates were present and reduced filamentation [20,22,23]. It also boosted lipid biosynthesis from fructose in an obese Y. lipolytica strain up to 55% [24]. The overexpression had, however, no effect on preferential uptake of glucose over fructose. The preference might in turn result from the kinetic properties of YlHxk1 as the enzyme has higher affinity for glucose than for fructose with a Km of 0.38 mM and 3.56 mM, respectively [21], even more pronounced than for the ScHxk2 hexokinase of S. cerevisiae (0.25 mM for glucose; 1.5 mM for fructose). In contrast, hexokinase from Schizosaccharomyces pombe (SpHxk1) is fructophilic with a Km of 8.4 mM for glucose and 1.5 mM for fructose [25]. In this study we investigated the role of hexose transporters and hexokinase in the preferential uptake of glucose over fructose in Y. lipolytica. We first examined the kinetic properties of Yht1, Yht3 and Yht4 in a heterologous host and then, through overexpression of hexose transporters in combination with native YlHxk1 or heterologous SpHxk1, we showed rearrangement of glucose and fructose utilization profiles. We further indicated that the increased carbon flux can be directed to different metabolic pathways by changing cultivation conditions. In addition, we propose that the introduced modifications can be harnessed for production of polyols by Y. lipolytica grown in hexose-containing media.

2. Results 2.1. Expression of Y. lipolytica Hexose Transporters in Heterologous Host Induce Diverse Monosaccharide Utilization Phenotypes The heterologous host S. cerevisiae EBY.VW4000 devoid of hexose transporters [26] is well-suited for comparing hexose uptake characteristics conferred by single transporters and has been previously used for this purpose [20,27,28]. The S. cerevisiae strains expressing YHT1, YHT3H222 (from Y. lipolytica H222, named hereafter YHT3) or YHT4 were cultivated in media containing glucose, fructose, mannose and a mixture of glucose and fructose. To monitor the consumption of each monosaccharide, substrate concentrations were Int. J. Mol. Sci. 2021, 22, 9282 3 of 20 measured throughout the 48 hgrowth. Single sugar experiments pointed out different efficiencies for sugar uptake among the three transporters (Figure 1A). During the 48 h of cultivation, only the strain express- ingin media YHT3 containing completely glucose, consumed fructose, each mannosesubstrate. and This a strain mixture grew of glucose slightly and better fructose. on man- To nosemonitor than the on consumption glucose and offructose. each monosaccharide, In turn, S. cerevisiae substrate with concentrations YHT1 utilized were fructose measured more efficiently,throughout while the 48 the hgrowth. YHT4 strain exhibited a clear preference towards glucose. Both of the strainsSingle, however, sugar experiments preferred mannose pointed over out different glucose efficienciesand fructose, for and sugar these uptake differences among werethe three also transporters visible in the (Figure growth1A). profiles. During Strikingly, the 48 h of the cultivation, YHT1-expressing only the strain was expressing unable YHT3to completelycompletely utilize consumed any of the each available substrate. substrates This strain within grew 48 h slightly. better on mannose thanWhen on glucose grown and in a fructose.medium containing In turn, S. a mixture cerevisiae ofwith glucoseYHT1 andutilized fructose, fructose all the strains more showedefficiently, a whileclear glucose the YHT4 preference strain exhibited (Figure a clear1B). Only preference the YHT3 towards-expressing glucose. strain Both of com- the pletelystrains, utilized however, both preferred sugars mannosewithin the over given glucose timeframe. and fructose, In contrast, and thesethe YHT1 differences- and YHT4 were- expressalso visibleing strains in the growthdid not profiles.consume Strikingly,fructose atthe all,YHT1 and the-expressing YHT4 strain strain used was glucose unable at to a 2.6completely-fold higher utilize rate any than of the the YHT1 available strain. substrates within 48 h.

Figure 1. Sugar utilization and growth of S. cerevisiae EBY.VW4000 transformants expressing hexose transporters from from Y.Y. lipolyticalipolytica in singlesingle sugarsugar mediamedia (A) andand mediummedium containingcontaining a mixturemixture ofof glucoseglucose and fructosefructose (B).). TheThe cultivationscultivations werewere carried out in three biologicalbiological replicates.replicates.

When grown in a medium containing a mixture of glucose and fructose, all the strains showed a clear glucose preference (Figure1B). Only the YHT3-expressing strain completely utilized both sugars within the given timeframe. In contrast, the YHT1- and YHT4-expressing strains did not consume fructose at all, and the YHT4 strain used glucose at a 2.6-fold higher rate than the YHT1 strain.

2.2. Construction and Growth Proﬁling of Y. lipolytica Strains Overexpressing Hexose Transporters and Hexokinases The different characteristics of sugar uptake observed in S. cerevisiae EBY.VW4000 with Yht1, Yht3 and Yht4 encouraged us to examine the behavior of Y. lipolytica expressing them individually and in combinations. Previously engineered yht1-4∆ mutant of Y. Int. J. Mol. Sci. 2021, 22, 9282 4 of 20

lipolytica W29 (JMY4788) impeded in hexose transport was used as the parental strain. Transporters were associated with two alternative hexokinases: the native YlHxk1 and SpHxk1 from S. pombe known for its higher afﬁnity for fructose. First, expression cassettes containing the pTEF-driven YlHXK1 and SpHXK1 genes were introduced into YHT1-4∆ mutant. From these two strains, derivatives with the desired combinations of transporters were developed. The YHT genes were placed under control of strong, constitutive pTEF promoter (see Section4 for details). In order to avoid artefacts caused by random integration of the cassettes in the genome, three clones for each genotype were analyzed for their growth ability in media containing glucose, fructose and mixture of both sugars. In the microcultivation experiments the transformants of the same genotype exhibited similar growth proﬁles, presented as average growth curves in Figure S1. All strains were able to grow in glucose and/or fructose, however, with varying rates. Thus, YHT4 caused poor growth in fructose when associated with SpHXK1 in the strains SpH1-Y4 and SpH1-Y1-Y4 (Figure S1D). Major variations were observed with YHT1, which conferred slow growth in fructose, in association with either hexokinase. Co-expression of both YHT1 and YHT4 in the SpH1 strain enabled weak growth in fructose (Figure S1). Only subtle differences were observed for the remaining transformants.

2.3. Rearrangement of Sugar Utilization Profile through Overexpression of Hexose Transporters in Combination with Hexokinases in Y. lipolytica To analyze their sugar preference, transformants were further characterized for their substrate consumption profiles during their growth in flasks containing a minimal medium supplemented with glucose and fructose as a mixture. Two strains of each genotype were used in the following experiment. Each genotype conferred the same phenotypic changes; therefore, data for one strain of each genotype is presented. All of the constructed strains used glucose prior to fructose; however, differences in the consumption rates of individual hexoses were observed. In the background of wild-type transporter genes at their original loci, the native YlHXK1 (control strain JMY2900) as well as the additional pTEF-YlHXK1 and pTEF-SpHXK1 cassettes resulted in similar consumption profiles (Figure2A); glucose was depleted around 30 h of culture, fructose uptake started at about 24–27 h and was completed at around 50 h. Overexpressing the transporter genes, in any combination along with the hexokinases, accelerated hexose consumption, leading to glucose and fructose depletion within 15–21 h and 27–33 h, respectively. One striking exception was observed with the YHT3 transporter, which the effect of strongly varied depending on the hexokinase. Overexpression of both YHT3 and YlHXK1 resulted in an early onset of fructose utilization, at around 12 h of culture (Figure2B). Despite that, the initial consumption rate was inconsiderable and increased only when glucose was completely exhausted and ceased at around 36 h. In contrast, the co-expression of SpHXK1 and YHT3 greatly improved utilization of both glucose and fructose. The sugars were consumed nearly in parallel from as early as 6 h of culture and were completely depleted within 30 h. This result places the SpH1-Y3 strain among the best ones for hexose uptake. The difference between both strains was also clear in their growth profiles, as the YlH1-Y3 mutant showed slower growth, reaching stationary phase 18 h later than its SpHXK1- expressing counterpart. Interestingly, when either YHT1 or YHT4 were expressed with YlHXK1 and YHT3, the growth and substrate utilization improved (Figure2C ). The same transporter complementation in the SpHXK1–YHT3 background did not much modify its phenotype. On the other hand, expression of the three transporters with either of the hexokinases resulted in a similar phenotype, close to that observed for other strains expressing SpHXK1 and YHT3 (Figure2D). Interestingly, the maximum glucose utilization rates were decreased in the strains expressing three transporters (Figure2E ). Furthermore, fructose was consumed 1.4-fold faster by the YlH1-Y1-Y3-Y4 strain than by the SpHXK1- expressing mutant with the same set of transporters, while glucose was consumed at the same rate. The remaining eight strains did not exhibit distinct differences in terms of growth and substrate utilization (Figure S2). Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 5 of 20

for other strains expressing SpHXK1 and YHT3 (Figure 2D). Interestingly, the maximum glucose utilization rates were decreased in the strains expressing three transporters (Fig- ure 2E). Furthermore, fructose was consumed 1.4-fold faster by the YlH1-Y1-Y3-Y4 strain Int. J. Mol. Sci. 2021, 22, 9282 than by the SpHXK1-expressing mutant with the same set of transporters, while glucose5 of 20 was consumed at the same rate. The remaining eight strains did not exhibit distinct differences in terms of growth and substrate utilization (Figure S2).

FigureFigure 2. 2. Growth and and substrate substrate utilization utilization profiles profiles of ofthe the Y. lipolyticaY. lipolytica transformantstransformants expressing expressing com- combinationsbinations of hexokinases of hexokinases and and hexose hexose transporters transporters in flasks in flasks in a in defined a defined medium medium containing containing a mix- a mixtureture of glucose of glucose and and fructose fructose (A–D (A).– MaximumD). Maximum glucose glucose and fructose and fructose utilization utilization rates ( ratesE). Red (E ).vertical Red dashed lines indicate onset of fructose utilization. The cultivations were performed in three biolog- vertical dashed lines indicate onset of fructose utilization. The cultivations were performed in three ical replicates. biological replicates.

InIn sum, sum, the the identity identity of of both both transporter transporter and and hexokinase, hexokinase, had had strong strong impact impact on on sugar sugar utilizationutilization proﬁles. profiles. All All the the strains strains with withSpHXK1 SpHXK1and andYHT3 YHT3overexpressed overexpressed and and only only them them startedstarted to to consume consume fructose fructose very very early, early, in in the the time time when when glucose glucose had had hardly hardly begun begun to to be be consumed.consumed. In In the the following following step, step, we we were were interested interested to to examine examine the the behavior behavior of of the the strains strains inin the the strictly strictly controlled controlled bioreactor bioreactor conditions. conditions. In In particular, particular, we we wondered wondered whether whether the the rapidrapid and and parallel parallel substrate substrate utilization utilization by by the the strains strains expressing expressingSpHXK1 SpHXK1and andYHT3 YHT3could could outperform the other strains in a biotechnological process and lead to its shortening. For this purpose, we chose the process of citric acid (CA) production, well-described for Y. lipolytica.

2.4. The Input Flux of Hexoses Impacts Downstream Metabolic Pathways For strain comparisons in the process of CA production, the cultivations were carried out in a deﬁned medium optimized for CA biosynthesis containing a mixture of glucose Int. J. Mol. Sci. 2021, 22, 9282 6 of 20

and fructose under carbon excess and nitrogen limitation in 5 L stirred-tank bioreactors. In general, sugar consumption proﬁles were similar to that described in the previous section. The rates of CA production and ﬁnal titers depended greatly on HXK/YHT genotypes (Table S2). The 17 bioreactor fermentations (one strain of each genotype) were performed in triplicates and the results are summarized in supplementary Figure S3. The major phenotypes are depicted in Figure3 and discussed below. The cultures were carried out Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 7 of 20 until both substrates were completely utilized or until 132 h in the case of the JMY2900, YlH1-Y3 and YlH1-Y3-Y4 strains.

FigureFigure 3. 3.Sugar Sugar consumption consumption and and citric citric acid acid (CA) (CA) secretion secretion of Y.of lipolyticaY. lipolyticastrains strains grown grown in a mediain a media optimized optimized for citric for citric acid productionacid production in 5 L in stirred 5 L stirred tank bioreactorstank bioreactors (A–C );(A metabolic–C); metabolic proﬁles profiles of the of analyzed the analyzed strains strains (D). Red (D). vertical Red vertical dashed dashed lines lines indicate the onset of fructose utilization. The cultivations were performed in three biological replicates. indicate the onset of fructose utilization. The cultivations were performed in three biological replicates.

Table 1. Parameters of bioreactorOverexpression cultures for citric of acid either production hexokinase using in Y. the lipolytica native transformantsYHT background in defined accelerated medium con- with a mixture of glucose and fructose. The cultivations were carried out until complete depletion of the substrates or sumption of both hexoses and increased CA production 2.3- (p = 6.2 × 10−6) and 2.7-fold until 132 h (strains JMY2900, YlH1-Y3 and YlH1-Y3-Y4) and were performed in three biological replicates. (p = 1.0 × 10−6; 67.9 g/L and 81.1 g/L vs. 29.4 g/L) for the YlH1 and SpH1 strain, re- YlH1-Y1- YlH1-Y3- YlH1-Y1- SpH1-Y1- SpH1-Y3- SpH1-Y1- Parameter Unit JMY2900 YlH1spectively SpH1 (Figure YlH13A).-Y3 These two strains showed the highestSpH1-Y3 CA yields from glucose and fructose among the analyzedY3 transformants Y4 (TableY3-Y4 1). A 24 h earlierY3 onsetY4 of fructoseY3-Y4 rxmax 0.13 0.14utilization 0.08 was observed0.10 for0.07 the YlH10.08 strain. The0.24 control0.17 JMY2900 0.19 strain 0.18 was unable0.22 to rglc g/L/h 0.73 1.04consume 1.03 all the available1.09 fructose1.26 within1.86 132 h1.35 of culture.1.16 In turn,1.53YHT3 in2.51 combination 1.35 rfru 0.43 0.84 0.77 0.41 0.91 0.89 1.77 1.34 1.43 1.70 1.67 with YlHXK1 had a detrimental effect on sugar consumption and CA biosynthesis when X 17.4 ± 1.2 14.5 ± 1.0 10.5 ± 0.9 13.5 ± 0.3 8.0 ± 1.0 10.5 ± 1.4 17.2 ± 2.4 11.9 ± 0.8 13.0 ± 1.3 12.6 ± 1.2 14.0 ± 1.8 compared to the YlH1 with native transporters70.19 ± 36.59 (Figure ± 3B). Conversely, the expression CA 29.39 ± 2.33 67.90 ± 3.51 81.09 ± 4.55 26.45 ± 1.24 65.50 ± 4.33 42.84 ± 1.83 41.87 ± 1.97 35.78 ± 1.76 33.19 ± 2.40 of an additional transporter (YHT1) restored2.22 the4.22 phenotype of the YlH1 strain, while the Lipids 5.74 ± 1.35 4.42 ±introduction 1.00 4.02 ± 0.98 of YHT43.23 ± 1.33 resulted 3.46 ± in1.03 an 3.61 intermediate ± 1.22 3.97 ± phenotype1.27 2.83 ± 0.66 of 3.51 the ± parental 0.83 3.51 ± and 0.97 YlHXK13.08 ± 1.56- Man g/L ND 1.34 ±expressing 0.27 ND strains.ND When threeND transporters 1.93 ± 0.06 were 8.47 ± overexpressed 1.03 8.47 ± 0.98 13.26 alongside ± 0.56 13.56YlHXK1 ± 0.96 13.42, sugars ± 1.88 Ara ND 0.90 ±were 0.03 completelyND exhaustedND withinND 0.78 72 h,± 0.13 but 4.98 the ± CA0.52 levels4.10 ± 0.08 showed 3.02 ± 0.22 a two-fold 4.08 ± 1.03 decrease3.20 ± 0.06 Ery ND NDwhen comparedND toND the mutantsND with twoND transporters 7.87 ± 1.12 (Figure11.94 ± 1.163B). 9.03 Nevertheless, ± 1.09 13.84 ± 1.69 due 9.34 to ±im- 0.97 Sum of ND 2.24proved sugarND consumptionND characteristics,ND 2.71 the CA21.32 productivities 24.51 were25.31 largely31.48 comparable 25.96 polyols to that of the YlH1 strain (Table1). YCA/S g/g 0.427 0.728 0.804 0.420 0.667 0.751 0.398 0.447 0.438 0.361 0.353 QCA g/L/h 0.111 0.314 0.307 0.100 0.287 0.266 0.254 0.298 0.303 0.248 0.263 qCA g/g/h 0.013 0.043 0.060 0.015 0.072 0.051 0.030 0.050 0.047 0.039 0.038 Abbreviations: Ara—arabitol; CA—citric acid; Ery—erythritol; Man—mannitol; ND—not detected; QCA—CA volumetric productivity; qCA—CA specific productivity; rfru—maximum fructose utilization rate; rglc—maximum glucose utilization rate; rxmax—maximum growth rate; S—sum of glucose and fructose concentrations at the beginning of cultivation; X—biomass; Y—yield.

2.5. Overexpression of a Fructophilic Hexokinase and Three Hexose Transporters Induces Rapid Biosynthesis of Polyols from Glucose and Fructose Significant amounts of polyols produced by the Y. lipolytica strains overexpressing SpHXK1 and YHT3 in various combinations held promise that their secretion might be further enhanced by optimizing culture conditions. To verify this, the strain SpH1-Y1-Y3-

Int. J. Mol. Sci. 2021, 22, 9282 7 of 20

Table 1. Parameters of bioreactor cultures for citric acid production using Y. lipolytica transformants in deﬁned medium with a mixture of glucose and fructose. The cultivations were carried out until complete depletion of the substrates or until 132 h (strains JMY2900, YlH1-Y3 and YlH1-Y3-Y4) and were performed in three biological replicates.

YlH1-Y1- YlH1-Y3- YlH1-Y1- SpH1-Y1- SpH1-Y3- SpH1-Y1- Parameter Unit JMY2900 YlH1 SpH1 YlH1-Y3 SpH1-Y3 Y3 Y4 Y3-Y4 Y3 Y4 Y3-Y4

rxmax 0.13 0.14 0.08 0.10 0.07 0.08 0.24 0.17 0.19 0.18 0.22 rglc g/L/h 0.73 1.04 1.03 1.09 1.26 1.86 1.35 1.16 1.53 2.51 1.35 rfru 0.43 0.84 0.77 0.41 0.91 0.89 1.77 1.34 1.43 1.70 1.67 X 17.4 ± 1.2 14.5 ± 1.0 10.5 ± 0.9 13.5 ± 0.3 8.0 ± 1.0 10.5 ± 1.4 17.2 ± 2.4 11.9 ± 0.8 13.0 ± 1.3 12.6 ± 1.2 14.0 ± 1.8 CA 29.39 ± 2.33 67.90 ± 3.51 81.09 ± 4.55 26.45 ± 1.24 65.50 ± 4.33 70.19 ± 2.22 36.59 ± 4.22 42.84 ± 1.83 41.87 ± 1.97 35.78 ± 1.76 33.19 ± 2.40 Lipids 5.74 ± 1.35 4.42 ± 1.00 4.02 ± 0.98 3.23 ± 1.33 3.46 ± 1.03 3.61 ± 1.22 3.97 ± 1.27 2.83 ± 0.66 3.51 ± 0.83 3.51 ± 0.97 3.08 ± 1.56 Mang/L ND 1.34 ± 0.27 ND ND ND 1.93 ± 0.06 8.47 ± 1.03 8.47 ± 0.98 13.26 ± 0.56 13.56 ± 0.96 13.42 ± 1.88 Ara ND 0.90 ± 0.03 ND ND ND 0.78 ± 0.13 4.98 ± 0.52 4.10 ± 0.08 3.02 ± 0.22 4.08 ± 1.03 3.20 ± 0.06 Ery ND ND ND ND ND ND 7.87 ± 1.12 11.94 ± 1.16 9.03 ± 1.09 13.84 ± 1.69 9.34 ± 0.97 Sum of ND 2.24 ND ND ND 2.71 21.32 24.51 25.31 31.48 25.96 polyols

YCA/S g/g 0.427 0.728 0.804 0.420 0.667 0.751 0.398 0.447 0.438 0.361 0.353 QCA g/L/h 0.111 0.314 0.307 0.100 0.287 0.266 0.254 0.298 0.303 0.248 0.263 qCA g/g/h 0.013 0.043 0.060 0.015 0.072 0.051 0.030 0.050 0.047 0.039 0.038

Abbreviations: Ara—arabitol; CA—citric acid; Ery—erythritol; Man—mannitol; ND—not detected; QCA—CA volumetric productivity; qCA—CA speciﬁc productivity; rfru—maximum fructose utilization rate; rglc—maximum glucose utilization rate; rxmax—maximum growth rate; S—sum of glucose and fructose concentrations at the beginning of cultivation; X—biomass; Y—yield. Int. J. Mol. Sci. 2021, 22, 9282 8 of 20

YHT3 combined with SpHXK1 clearly accelerated utilization rates of glucose and fructose and this effect was consistent regardless of additional transporters (Figure3C). The strain SpH1-Y3 as well as SpH1-Y1-Y3 showed similar, yet noticeably, decreased CA secretion when compared to the SpH1 strain. Transformants expressing SpHxk1 in combination with YHT3 and YHT4, only YHT4 or all three transporters had even lower CA biosynthetic abilities (Figure3C and Figure S3). The obtained data showed that the observed decrease was accompanied by an increased polyol secretion of which mannitol and erythritol were predominant (Figure3D, Table1).

2.5. Overexpression of a Fructophilic Hexokinase and Three Hexose Transporters Induces Rapid Biosynthesis of Polyols from Glucose and Fructose Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 8 of 20 Signiﬁcant amounts of polyols produced by the Y. lipolytica strains overexpressing SpHXK1 and YHT3 in various combinations held promise that their secretion might be further enhanced by optimizing culture conditions. To verify this, the strain SpH1-Y1- Y4Y3-Y4 was was chosen chosen as it asshowed it showed good good polyol polyol production production accompanied accompanied by nearly by nearly parallel parallel con- sumptionconsumption of glucose of glucose and andfructose fructose and andthe fastest the fastest fermentation fermentation time time(Figure (Figure 3C). 3TheC). Themu- tantmutant waswas cultivated cultivated in a inmedia a media optimized optimized for polyol for polyol biosynthesis biosynthesis containing containing either either a mix- a turemixture of glucose of glucose and and fructose fructose or orglycerol glycerol as asa carbon a carbon source source (Figure (Figure 44).). In In the sugar-basedsugar-based medium,medium, the the substrates substrates were were rapidly utilized and completely exhausted within 48 h.h. The strainstrain secrete secretedd 25 25-fold-fold less less CA CA than than in in the the CA CA-optimized-optimized media but instead 1.5 1.5-fold-fold more polyols (25.96 g/L vs. vs. 40.58 40.58 g/L) g/L) of of which which 44% 44% mannitol, mannitol, 35% 35% erythritol erythritol and and 21% 21% arabitol (Table(Tables2 2and and Table S3). Mannitol S3). Mannitol production production accelerated accelerated significantly signiﬁcantly with the with onset the onsetof linear of frlineaructose fructose consumption consumption and lasted and lastedbetween between 24 and 24 34 and h of 34culture h of culture (Figure (Figure 4A). In4 A).turn, In when turn, glycerolwhen glycerol was used, was a used, different a different quantitative quantitative polyol polyolspectrum spectrum was obtained was obtained in which in whicheryth- ritolerythritol was predominant was predominant (Figure (Figure 4B). 4LowB). Low amounts amounts of mannitol, of mannitol, arabitol arabitol and and CA CAwere were se- cretedsecreted while while the the biomass biomass was was comparable comparable to to that that in in the the sugar sugar-based-based medium. Glycerol was used within 58 h h..

Figure 4. Substrate utilization, growth, growth, citric acid acid (CA) (CA) and and polyol polyol production production by by the the strain strain Y. lipolytica SpH1SpH1-Y1-Y3-Y4-Y1-Y3-Y4 in 5 L stirredstirred-tank-tank bior bioreactorseactors in media optimized for polyol production containing a mixture of glucose glucose and and fructose ( (A) or glycerol (B); in media optimized for erythritol production containing a mixture of glucose and fructose (C) or glycerol (D). glycerol (B); in media optimized for erythritol production containing a mixture of glucose and fructose (C) or glycerol (D). The cultivations were performed in three biological replicates. The cultivations were performed in three biological replicates. The SpH1-Y1-Y3-Y4 mutant was additionally tested in the media containing NaCl The SpH1-Y1-Y3-Y4 mutant was additionally tested in the media containing NaCl which was applied in order to increase osmotic pressure and promote erythritol biosyn- which was applied in order to increase osmotic pressure and promote erythritol biosynthesis. In the glucose- and fructose-based media the strain consumed available sugars thesis. In the glucose- and fructose-based media the strain consumed available sugars within 48 h and produced 23.12 g/L erythritol, 14.66 g/L arabitol, negligible amounts of within 48 h and produced 23.12 g/L erythritol, 14.66 g/L arabitol, negligible amounts mannitol and no CA throughout (Figure 4C). In turn, when grown on glycerol, the trans- of mannitol and no CA throughout (Figure4C). In turn, when grown on glycerol, the formant needed 36 h more to entirely utilize the substrate. Glycerol clearly favored eryth- transformant needed 36 h more to entirely utilize the substrate. Glycerol clearly favored ritol biosynthesis at an expense of decreased arabitol production; however, in the final erythritol biosynthesis at an expense of decreased arabitol production; however, in the stageﬁnal stageof the of process, the process, CA secretion CA secretion occurred occurred (Figure (Figure 4D).4 D).Compared Compared to the to thecultivation cultivation on monosaccharides, mannitol was not detected and nearly 20% less biomass was made (Ta- ble 2). Interestingly, maximum glucose and fructose utilization rates were, respectively, 1.1- and 1.3-fold higher in the erythritol medium than in the polyol biosynthesis broth. Conversely, glycerol was consumed with a 2.1-fold slower rate in the erythritol medium when compared to the polyol medium.

Int. J. Mol. Sci. 2021, 22, 9282 9 of 20

on monosaccharides, mannitol was not detected and nearly 20% less biomass was made (Table2). Interestingly, maximum glucose and fructose utilization rates were, respectively, 1.1- and 1.3-fold higher in the erythritol medium than in the polyol biosynthesis broth. Conversely, glycerol was consumed with a 2.1-fold slower rate in the erythritol medium when compared to the polyol medium.

Table 2. Parameters of bioreactor cultures for polyol biosynthesis using Y. lipolytica SpH1-Y1-Y3-Y4 in media containing a mixture of glucose and fructose or glycerol. The cultivations were carried out until complete depletion of the substrate(s) and were performed in three biological replicates.

Polyol Media Erythritol Media Parameter Unit Glucose Fructose Glycerol Glucose Fructose Glycerol

rxmax 0.693 0.857 0.767 0.400 r 2.048 - 2.738 - glc g/L/h rfru 2.698 - 3.229 - rgly - 2.432 - 2.013 Man 18.03 ± 1.37 5.04 ± 0.90 2.09 ± 0.66 0.11 ± 0.01 Arag/L 8.48 ± 0.68 4.29 ± 1.02 14.66 ± 1.06 4.07 ± 0.32 Ery 14.07 ± 1.29 24.28 ± 0.99 23.12 ± 1.43 28.31 ± 1.33

YMan/S 0.199 0.049 0.021 0.001 YAra/S 0.094 0.041 0.145 0.037 Y 0.155 0.235 0.229 0.257 Ery/S g/g YMan/X 1.050 0.319 0.101 0.007 YAra/X 0.493 0.271 0.708 0.238 YEry/X 0.819 1.534 1.117 1.655

qMan 0.022 0.012 0.002 0.000 qAra g/g/h 0.010 0.006 0.015 0.003 qEry 0.017 0.018 0.023 0.020

QMan 0.188 0.043 0.022 0.000 QAra g/L/h 0.088 0.037 0.153 0.024 QEry 0.147 0.209 0.241 0.169 Abbreviations: Ara—arabitol; Ery—erythritol; Man—mannitol; Q—volumetric productivity; q—CA speciﬁc productivity; rfru—maximum fructose utilization rate; rglc—maximum glucose utilization rate; rgly—maximum glycerol utilization rate; rxmax—maximum growth rate; S—sum of substrate concentrations at the beginning of cultivation; X—biomass; Y—yield.

3. Discussion In the Y. lipolytica genome, there are six genes encoding hexose transporters (YHT1 to YHT6) of which YHT1, YHT3 and YHT4 code for efficient broad-range transporters enabling uptake of glucose, fructose, mannose and galactose [20]. The YHT1 and YHT4 genes were previously shown to be consistently transcribed in media containing glucose and fructose, and are, therefore, deemed the main hexose transporter-encoding genes [20]. On the other hand, YHT3 expression was observed mostly in the stationary phase and occasionally in the exponential growth phase [20], suggesting an auxiliary role of this transporter in sugar uptake. In our study, we used previously constructed S. cerevisiae strains expressing YHT1, YHT3 (YHT3H222 allele from Y. lipolytica H222) and YHT4 [20], based on the EBY.VW4000 strain devoid of all 17 hexose transporters [26] to determine the kinetic characteristics of each transporter. To our surprise, YHT3 turned out to be the most efficient one when expressed in a heterologous host, enabling rapid utilization of glucose, fructose and mannose. It is, therefore, interesting that its expression was previously observed mainly in the stationary growth phase in Y. lipolytica [20]. Furthermore, all three transporters showed clear preference towards mannose which is not without significance for future studies on utilization of mannose-containing lignocellulosic biomass in bioprocesses using Y. lipolytica. Microbial growth from a carbon source results from both transport and downstream metabolic pathways for energy and metabolite production. As previously reported, phos- Int. J. Mol. Sci. 2021, 22, 9282 10 of 20

phorylation of fructose by hexokinase for entering glycolysis (Figure5) is a limiting step in the Y. lipolytica wild-type W29 strain [24]. In order to examine the effect of individual transporters on hexose utilization by Y. lipolytica, we used the YHT1-4∆ mutant of W29 (strain JMY4788) impeded in hexose transport, unable to grow on glucose- nor on fructose-containing solid media (Figure S4). We previously constructed an overexpression cassette by placing the native hexokinase YlHXK1 gene under the strong and constitutive pTEF promoter (pTEF-YlHXK1)[24]. As some hexokinases exhibit higher affinity towards fructose than for glucose, such as the SpHxk1 [25], an expression cassette with the pTEF-driven SpHXK1 was constructed. We generated strains bearing either the YlHXK1 or the SpHXK1 overexpression cassette in the wild-type derivative strain (JMY2101) and in the auxotrophic YHT1-4∆ mutant derivative (PHY130) additionally overexpressing each of the three essential transporters, namely YHT1, YHT3H222 and YHT4 individually, in combinations of two and in a combination of three transporters. Previous experiments showed that wild-type Y. lipolytica strains grow poorly on fructose, but this can be overcome by hexokinase overexpression [24]. The introduction of heterologous hexokinase exerted the same effect as evidenced here with SpHxk1 or previously with ScHxk2 and YayaHxk1 [22]. The interplay between hexokinase and hexose transporters, however, seem to be crucial in the monosaccharide metabolism, especially fructose, in Y. lipolytica as clear differences in growth profiles, depending on the particular transporters associated with either hexokinase, were observed. The true nature of these interactions cannot be inferred using only growth profiles. We hypothesized that the molecular mechanism controlling preferential uptake of glucose over fructose in Y. lipolytica might hold in the affinity toward glucose and fructose of hexose transporters and/or hexokinase. YlHxk1 has over 9-fold higher affinity for glucose than for fructose [21], which could be one of the reasons for the yeasts’ strict glucose preference. As mentioned above, some yeast species, such as S. pombe, have fructophilic hexokinase with a higher affinity for fructose than for glucose [25]. In our study, overexpression of SpHxk1 did not change the substrate utilization pattern as glucose was always utilized first. This, however, might have occurred due to the existence of an intact glucokinase that could suffice for efficient glucose processing. The interplay between hexokinase and sugar transporter(s) was relevant, as, for example, YHT3 exerted varying effects on fructose utilization depending on the co-expressed hexokinase. Thus, when combined with YlHxk1, the onset of fructose consumption took place only after glucose depletion. Furthermore, fructose was imported to the cells until its concentration reached a threshold concentration of ca. 1.5 g/L in the shake-flask experiment, after which it was consumed with a very slow rate. In turn, YHT3 co-expressed with SpHxk1 triggered an entirely different phenotype allowing for rapid, nearly parallel utilization of both sugars to their complete exhaustion. In the SpHxk1 strains, the strong effect of YHT3 was present, while in the YlHxk1 background, additional transporters allowed for complete utilization of fructose, but in none of them parallel consumption of sugars was observed. The polarized, hexokinase-dependent effect of YHT3 suggests the interplay between YHT3 and hexokinase. It may be possible that the native Hxk1, which has a low affinity for fructose, directly or indirectly affects fructose transport through YHT3. On the other hand, the low Km for fructose of SpHxk1 allows for rapid fructose conversion. Furthermore, SpHxk1 might not interact with native proteins in the same manner as YlHxk1 does, therefore, allowing for uninterrupted fructose import. Structurally, YlHxk1 has a unique feature, namely a 37-amino acid loop not found in other known hexokinases, including SpHxk1 (Figure S5; [21]), which has been shown to be important in inhibitor binding and gene expression regulation [22]. This structure could, therefore, interact with YHT3 in effect decelerating sugar intake and defending the cell from the carbon overflow in the central carbon metabolism (Figure5) and/or increased intracellular osmotic pressure. The efficiency of substrate(s) utilization remains an important aspect in bioprocess design. Considering the rapid consumption of glucose and fructose by the SpHXK1 Int. J. Mol. Sci. 2021, 22, 9282 11 of 20

and YHT3-expressing strains, we were interested to see how they would perform in a biotechnological process. As a model bioprocess, we chose the extensively studied in Y. lipolytica process of citric acid (CA) production [19,29]. In concert with the results obtained from the shake-flask experiments, overexpression of either hexokinase accelerated fructose utilization. It also led to higher CA secretion, most likely caused by debottlenecked glycolysis (Figure5), as hexokinase is one of the major factors limiting this core pathway in Y. lipolytica [23,24]. The strains expressing transporters alongside SpHxk1 utilized glucose and fructose more rapidly than their YlHxk1 counterparts and generally secreted less CA. Low CA yields of the YlH1-Y3 and control JMY2900 strain could be explained by incomplete fructose utilization in the given timeframe (132 h). As stated, the SpHxk1- expressing strains secreted much less CA than could be expected, what indicated that the available carbon was redirected to other pathways. Nitrogen limitation in the applied conditions is known to direct the carbon flux either towards CA biosynthesis or lipid accumulation [30]. The analyzed mutants did not differ much in the generated biomass and lipid content, they had, however, produced significant amounts of polyols, of which mannitol and erythritol were predominant. The increased polyol secretion may have three possible explanations: (1) hexose transporters in Y. lipolytica have a dual function—hexose import and polyol export; (2) the limited capacity of glycolysis results in the redirection of carbon flux to other pathways, such as pentose phosphate pathway (Figure5); (3) the high sugar influx results in rising osmotic pressure inside the cells, creating a need for a defense mechanism allowing for pressure equalization. While the first explanation is unlikely, as evidenced by the bioreactor experiment in which the YHT1-4∆ mutant secreted polyols (Figure S6), the second and third hypotheses are probable. Thus, carbon overflow through glycolysis can be compensated by directing the flux to pentose phosphate pathway, which results in increased erythritol biosynthesis. It would be also interesting to see, whether the elimination of more bottlenecks in glycolysis, for example by overexpression of PFK1 (Figure5), would push the carbon flux down the pathway to the TCA cycle instead of directing it towards pentose phosphates. At this point, it is worth mentioning that secreted CA can be re-incorporated into the metabolism when supplied carbon sources are depleted [19], which creates a need for careful bioprocess design. In turn, since it is known that Y. lipolytica secretes erythritol in response to the presence of glycerol [12], a similar mechanism might exist for sugar response. The rerouting of carbon excess toward polyol production led us to evaluate polyols production by our engineered strains. Specifically, we wondered whether the production capacities of the strains expressing SpHxk1 and YHT3 could be further improved by the optimization of culture conditions. For that purpose, we chose the SpH1-Y1-Y3-Y4 strain as it showed the fastest consumption of glucose and fructose accompanied by a significant secretion of polyols. The strain was cultivated in bioreactors in media optimized for polyol production [31], containing either a mixture of glucose and fructose or glycerol as a carbon source. In the conditions tested, Y. lipolytica SpH1-Y1-Y3-Y4 used glucose and fructose 24 h earlier than in the CA production media. At the same time, it secreted insignificant amounts of CA at an expense of increased polyol levels. For comparison, the strain was also cultivated on glycerol. The complete substrate exhaustion occurred 10 h later than in the hexose-based media and more CA was formed. The sum of produced polyols was lower; however, the obtained mixture was more homogenous as erythritol was predominant. The faster consumption of sugars in the optimized conditions could be explained by the higher availability of nitrogen in the medium, which translated into a higher growth rate [32]. In turn, glycerol could not be utilized as fast, as the strain did not contain any improvements in the glycerol assimilation pathway (Figure5); however, its utilization rate was nearly identical to that reported previously for the wild-type Y. lipolytica strain (0.02 mol/L/h vs. 0.022 mol/L/h) [32]. The obtained quantities of particular polyols strictly depended on the used substrates. Y. lipolytica tends to overproduce erythritol in response to the osmotic pressure induced by glycerol [12], and the strain used in our work behaved in the same manner. On the other hand, when grown on glucose and fructose, mannitol Int. J. Mol. Sci. 2021, 22, 9282 12 of 20

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 12 of 20 was predominant, followed by signiﬁcant amounts of erythritol and arabitol. In these conditions, increased mannitol biosynthesis could be expected as it is only one metabolic stepfrom from fructose fructose (Figure (Figure 5), carried5), carried out by out mannitol by mannitol dehydrogenase dehydrogenase [33,34], which [ 33,34 also], which coin- also coincidedcided with with the the period period of high of high-rate-rate fructose fructose consumption. consumption.

FigureFigure 5. Simpliﬁed 5. Simplified schematic schematic representation representation of of the the metabolic metabolic pathways for for polyol polyol and and CACA biosynthesis biosynthesis in the in yeast the yeastY. Y. lipolyticalipolytica. Abbreviations:. Abbreviations: CA—citric CA—citric acid; acid; DHAP—dihydroxyacetone DHAP—dihydroxyacetone phosphate; phosphate; E4P E4P—erythrose—erythrose 4-phosphate; 4-phosphate; F1.6BP F1.6BP—— fructose 1.6-bisphosphate; F6P—fructose 6-phosphate; FUM—fumaric acid; G3P—glycerol 3-phosphate; G6P—glucose 6- fructosephosphate; 1.6-bisphosphate; GA3P—glyceraldehyde F6P—fructose 3-phosphate; 6-phosphate; Gn6P FUM—fumaric—gluconolactone acid; 6-phosphate; G3P—glycerol ICA—isocitric 3-phosphate; acid; KGA G6P—glucose—α-ke- 6-phosphate;toglutarate; GA3P—glyceraldehyde MA—malic acid; OAA 3-phosphate;—oxaloacetate; Gn6P—gluconolactone PEP—phosphoenolpyruvate; 6-phosphate; PPP—pentose ICA—isocitric phosphate acid; pathway; KGA— α- ketoglutarate;R5P—ribose MA—malic 5-phosphate; acid; Ru5P OAA—oxaloacetate;—ribulose 5-phosphate; PEP—phosphoenolpyruvate; S7P—sedoheptulose 7-phosphate; PPP—pentose SUCC—succini phosphatec acid; Suc-pathway; R5P—riboseCoA—succinyl 5-phosphate;-CoA; TCA Ru5P—ribulose—tricarboxylic 5-phosphate; acid cycle; X5P S7P—sedoheptulose—xylulose 5-phosphate. 7-phosphate; Gene identifiers: SUCC—succinic ACO1 (YALI0D09361g acid; SucCoA—); AraDH1 (YALI0F02211g); AraDH2 (YALI0E05643g); CIT1 (YALI0E00638g); CIT2 (YALI0E02684g); ER (YALI0F18590g); FBA1 succinyl-CoA; TCA—tricarboxylic acid cycle; X5P—xylulose 5-phosphate. Gene identiﬁers: ACO1 (YALI0D09361g); (YALI0E26004g); FBP1 (YALI0A15972g); FUM1 (YALI0C06776g); GND1 (YALI0B15598g); GPD1 (YALI0B02948g); GUT1 AraDH1(YALI0F00484g (YALI0F02211g)); HXK1; AraDH2 (YALI0B22308g) (YALI0E05643g); IDH1 (YALI0E05137g; CIT1 (YALI0E00638g)); IDH2 (YALI0D06303g; CIT2 (YALI0E02684g)); IDP1 (YALI0F04095g; ER (YALI0F18590g)); KGD1 ; FBA1 (YALI0E26004g)(YALI0E33517g); KGD2; FBP1 (YALI0E16929g (YALI0A15972g)); LSC1; FUM1 (YALI0E24013g) (YALI0C06776g); LSC2 (YALI0D04741g); GND1 (YALI0B15598g); MaDH1 (YALI0D16753g); GPD1 (YALI0B02948g); MaDH2 ; GUT1 (YALI0F00484g)(YALI0E14190g); ;MDH2HXK1 (YALI0D18964g) (YALI0B22308g); ;PDA1IDH1 (YALI0F20702g (YALI0E05137g)); PFK1; IDH2 (YALI0D16357g (YALI0D06303g)); PGI1; IDP1 (YALI0F07711g (YALI0F04095g)); PYK1; KGD1 (YALI0F09185g); SDH1 (YALI0D11374g); SDH2 (YALI0D23397g); SDH3 (YALI0E29667g); SDH4 (YALI0A14784g); TAL1 (YALI0E33517g); KGD2 (YALI0E16929g); LSC1 (YALI0E24013g); LSC2 (YALI0D04741g); MaDH1 (YALI0D16753g); MaDH2 (YALI0F15587g); TKL1 (YALI0E06479g); TPI1 (YALI0F05214g); ZWF1 (YALI0E22649g). (YALI0E14190g); MDH2 (YALI0D18964g); PDA1 (YALI0F20702g); PFK1 (YALI0D16357g); PGI1 (YALI0F07711g); PYK1 (YALI0F09185g); SDH1 (YALI0D11374g)We were; SDH2 curious (YALI0D23397g) to whether ;theSDH3 polyol (YALI0E29667g) spectrum could; SDH4 be changed (YALI0A14784g) by further; opti-TAL1 (YALI0F15587g); TKL1 (YALI0E06479g)mization; TPI1 of culture (YALI0F05214g) conditions.; ZWF1 The (YALI0E22649g)optimization relied. on the addition of NaCl to the previously used media to increase the osmotic pressure and induce erythritol secretion We were curious to whether the polyol spectrum could be changed by further optimization of culture conditions. The optimization relied on the addition of NaCl to the

Int. J. Mol. Sci. 2021, 22, 9282 13 of 20

previously used media to increase the osmotic pressure and induce erythritol secretion [31]. Glucose and fructose were used as quickly as before, but more biomass and only trace amounts of CA were made. As expected, the polyol spectrum changed in favor of erythritol, and mannitol was produced in very low quantities. Recently, Wang et al. (2020) [34] carried out metabolic engineering of Y. lipolytica, aiming to increase erythritol biosynthesis through the identification and subsequent suppression of the genes involved in mannitol and arabitol formation. They pointed out that the knock-out of the gene MDH2 (YALI0D18964g) encoding mannitol dehydrogenase resulted in a complete loss of ability to produce mannitol, which accompanied a slight increase in erythritol production. We observed a similar phenomenon caused by the presence of salt in the medium. As evidenced by the gene expression data, NaCl (or increased osmotic pressure in general) caused a downregulation of MDH2 (Figure S7).Although the titers of polyols, especially erythritol obtained in our study, cannot compete with those previously reported, reaching or even exceeding 200 g/L during growth in glycerol-based media [12,35], we find it interesting to see that it is possible to obtain decent amounts of these compounds using hexoses found in abundance in industrial wastes or lignocellulosic materials [17,36]. To the best of our knowledge, we show here for the first time, rapid polyol production by Y. lipolytica in fructose-containing media through the modifications of the upstream cellular processes, i.e., sugar import and activation. We reason that further engineering of metabolic pathways directly implicated in polyol biosynthesis, such as pentose phosphate pathway (Figure5), could enhance the production capacities of the SpH1-Y1-Y3-Y4 strain, as it was previously reported in transketolase-(TKL1), erythrose reductase-(ER) or erythritol dehydrogenase-(EYD1) overexpressing Y. lipolytica strains [37–40]. As it is in the case of CA, secreted erythritol can also be catabolized by the Y. lipolytica cells when carbon sources are depleted, a process in which erythritol kinase encoded by the EYK1 gene plays a major role [41]. Therefore, careful bioprocess characterizations, in which the exact time points of substrate exhaustion are provided, as presented in the present study, and/or the disruption of EYK1 could lead to higher yields. From an industrial perspective, it would also be of great interest to analyze the performance of the SpH1-Y1-Y3-Y4 derivatives with overexpressed invertase [42] and modified downstream metabolic pathways in industrial setups, including the use of molasses.

4. Materials and Methods 4.1. Culture Media The Escherichia coli strains were maintained in a LB medium, containing 5 g/L yeast extract (YE), 10 g/L tryptone, 10 g/L NaCl with 20 g/L agar in plates and 0.05 mg/L kanamycin for selection of transformants at 37 ◦C. Y. lipolytica strains were maintained in a YPG medium, consisting of 10 g/L yeast extract, 20 g/L peptone, 20 g/L glycerol, 20 g/L agar (for plates) and 0.2 g/L Hygromycin B (when applicable) at 28 ◦C. Minimal (YNB) medium for selection of Y. lipolytica transformants was prepared using a 1.9 g/L yeast nitrogen base (without amino acids and ammonium sulfate; Formedium Ltd., Hunstanton, UK), 10 g/L glycerol, 5 g/L NH4Cl, 50 mM phosphate buffer pH 6.8 with 20 g/L agar and 0.1 g/L uracil or 0.2 g/L leucine when needed. For long-term storage the strains were kept in 500 g/L glycerol at −80 ◦C. Minimal media with single sugar used for S. cerevisiae and Y. lipolytica substrate utilization kinetics and Y. lipolytica microcultivations contained 1.9 g/L YNB, 5 g/L NH4Cl, 50 mM phosphate buffer pH 6.8 and 10 g/L glucose, fructose or mannose. The medium with a mixture of glucose and fructose used in microcultivations and S. cerevisiae and Y. lipolytica substrate utilization kinetics consisted of 1.9 g/L YNB, 10 g/L NH4Cl, 50 mM phosphate buffer pH 6.8, 10 g/L glucose and 10 g/L fructose. The medium for citric acid production contained 50 g glucose, 50 g fructose, 1.5 g −3 NH4Cl, 0.7 g KH2PO4, 1.0 g MgSO4·7H2O, 0.3 g YE and 0.3 × 10 g thiamine in 1 L of tap water. The media optimized for the production of polyols and erythritol were adopted from [31]. The polyol production media consisted of a mixture of 50 g glucose and 50 g Int. J. Mol. Sci. 2021, 22, 9282 14 of 20

fructose or 100 g glycerol as the carbon source, 2.7 g (NH4)2SO4, 0.2 g KH2PO4, 1.0 g MgSO4·7H2O, 1.6 g YE in 1 L tap water with 3.0 g CaCO3 (for the shake-ﬂask experiment). The erythritol media was composed of a mixture of 50 g glucose and 50 g fructose or 100 g glycerol, 4.6 g (NH4)2SO4, 0.22 g KH2PO4, 1.0 g MgSO4·7H2O, 1.0 g YE and 25.5 g NaCl in 1 L tap water with the addition of 3.0 g CaCO3 in the shake-ﬂask experiment. Stocks of sugars and glycerol were of concentration 500 g/L and cold-sterilized through 0.22 µm membranes. The remaining media components were autoclaved at 121 ◦C for 20 min.

4.2. Strain Construction The S. cerevisiae strains expressing hexose transporters from Y. lipolytica were constructed in our previous study [20]. The transporter and hexokinase genes from Y. lipolytica were amplified from a wild- type W29 strain genome using Phusion High-Fidelity DNA polymerase (Thermo Scientific, Waltham, MA, USA) and primers listed in Table S1 with overhangs containing BamHI (or BglII) and AvrII recognition sites. Amplified DNA fragments were then gel-purified using the Gel-Out Concentrator kit (A&A Biotechnology, Gdynia, Poland) and digested with appropriate restriction enzymes. Prepared fragments were then cloned into BamHI/AvrII digested and dephosphorylated JMP62-URA3ex-pTEF or JMP62-LEU2ex-pTEF vectors containing zeta sequences for random chromosomal integration [43]. The vectors were then used for transformation of competent E. coli DH5α strain. The colonies were verified using the PCR-colony technique and Sanger sequencing (Genomed S.A., Warsaw, Poland). Correct vectors were then digested with NotI to release expression cassettes and were used for Y. lipolytica transformations. All the Y. lipolytica transformations were performed using the lithium acetate method [44]. SpHXK1 from S. pombe was obtained as a synthetic gene codon-optimized for expression in Y. lipolytica (GenScript, Leiden, the Netherlands) containing BamHI/AvrII overhangs, cloned into JMP62-URA3ex-pTEF or JMP62-LEU2ex- pTEF vectors and used for Y. lipolytica transformations (Figure6, Table S4 [ 45,46]). The obtained transformants were analyzed using PCR on genomic DNA extracted using the Genomic Mini AX Yeast kit (A&A Biotechnology) with primers listed in Table S1. The nucleotide sequence of optimized SpHXK1 is given in Supplementary File S1. The deduced amino acid sequence is identical to the NCBI Reference Sequence NP_592948.1. The Y. lipolytica strains were constructed in an YHT1-4∆ genetic background in which native hexokinase and glucokinase were intact (Y. lipolytica JMY4788; [20]). The markers in this strain were rescued using CreLox recombinase [43]. Next, two hexokinases, native YlHXK1 (YALI0B22308g) and heterologous SpHXK1 were overexpressed, creating a starting point for two branches of subsequent transformants (Figure6). These two strains were further engineered by the overexpression of single (YHT1 (YALI0C06424g), YHT3 (YALI0F19184g), YHT4 (YALI0E23287g)), combinations of two (YHT1-YHT3, YHT1-YHT4, YHT3-YHT4) and three (YHT1-YHT3-YHT4) transporter genes using the NotI-digested JMP62-URA3ex-pTEF- or JMP62-LEU2ex-pTEF-based vectors. The obtained transformants were analyzed using PCR on genomic DNA. In each branch there was one step required for marker rescue and one for prototrophy restoration using the LEU2ex fragment obtained from JMP62-LEU2ex-pTEF plasmid via I-SceI digestion. Due to the randomness of the genome integration site and possible phenotypic variations resulting from the nature of the employed overexpression method, three strains of each genotype were used for the initial experiments. Int. J. Mol. Sci. 2021, 22, 9282 15 of 20 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 15 of 20

Figure 6. Construction of of Y. lipolytica strains expressing hexokinases and hexose transporters. A A detailed detailed list list of of the the Y.Y. lipolytica strains used in this study is available in Table S4. lipolytica strains used in this study is available in Table S4. The Y. lipolytica strains were constructed in an YHT1-4Δ genetic background in which native4.3. Microcultivations hexokinase and glucokinase were intact (Y. lipolytica JMY4788; [20]). The markers in thisFor strain growth were proﬁling rescued of usingY. lipolytica CreLoxtransformants recombinase [ growing43]. Next, on two glucose, hexokinases, fructose native and a YlHXK1mixture of(YALI0B22308g both in minimal) and YNB heterologous media a Synergy SpHXK1 H1 were microplate overexpressed, reader (BioTek, creating Winooski, a start- ingVT, point USA) for was two used. branches Prior of to subsequent the experiment transformants the cells were (Figure grown 6). These in 5 mL two YPG strains medium were furtherovernight engineered at 28 ◦C and by 180 the rpm, overexpression washed thrice with of single sterile (distilledYHT1 ( waterYALI0 andC06424g standardized), YHT3 (toYALI0F19184g OD10. Cultivations), YHT4 were(YALI0E23287g carried out)), in combinations 96-well microtiter of two plates(YHT1 (NEST,-YHT3, Wuxi,YHT1- China)YHT4, YHT3with a-YHT4 working) and volume three of (YHT1 200 µ-YHT3L. Initial-YHT4 OD) was transporter set to 0.5. genes Throughout using the the Not process,I-digested the JMP62temperature-URA3ex was-pTEF maintained- or JMP62 at 28 ◦-CLEU2ex and 600-pTEF rpm- linearbased shaking vectors was. The applied. obtained The growth trans- formantswas monitored were analyzed by measuring using opticalPCR on density genomic in DNA 10 min. In intervals. each branch The there experiments was one were step requiredconducted for in marker three biological rescue and replicates. one for prototrophy restoration using the LEU2ex fragment obtained from JMP62-LEU2ex-pTEF plasmid via I-SceI digestion. Due to the ran- domness4.4. Substrate of the Utilization genome integration Kinetics site and possible phenotypic variations resulting from the natureThe substrate of the employed utilization overexpression kinetics were carriedmethod, out three in 250 strains mL Erlenmeyerof each genotype ﬂasks were with used50 mL for YNB the mediuminitial experiments. with a mixture of glucose and fructose. Prior to inoculation, the cells were prepared as described in Section 4.3. Prepared media were inoculated to an OD 0.5

Int. J. Mol. Sci. 2021, 22, 9282 16 of 20

and incubated at 28 ◦C on a rotary shaker with 180 rpm shaking speed. The samples were taken in 3 h intervals for growth and substrate utilization analysis. The experiments were conducted in three biological replicates and standard deviations were calculated.

4.5. Bioreactor Cultivations The bioreactor cultivations were carried out in 5 L stirred-tank BIOSTAT B plus reactors (Sartorius AG, Frankfurt, Germany). The bioreactors with appropriate media were sterilized at 121 ◦C for 20 min. Carbon sources were cold-sterilized and added directly to the vessels after sterilization in autoclaves. Prior to inoculations, the cells were grown in 50 mL YPG medium in 250 mL Erlenmeyer ﬂasks at 28 ◦C, 180 rpm overnight and subsequently washed thrice with sterile distilled water. The initial OD was set to 0.5 and the working volume was 2 L. The cultivations were conducted at 28 ◦C with a stirring rate of 800 rpm and an aeration rate of 0.8 vvm (vessel volume per minute). pH was automatically maintained at 6.8 (for citric acid production) or 3 (for polyols) using 40% (w/v) NaOH solution. All the bioreactor experiments were conducted in three biological replicates and standard deviations were calculated.

4.6. Analytical Methods The optical density measurements in the kinetic experiments were carried out using the SmartSpec Plus spectrophotometer (Biorad, Hercules, CA, USA). The dry biomass during bioreactor cultures was determined gravimetrically after cells being filtered through 0.22 µm membranes and dried at 105 ◦C using laboratory scale with moisture analyzer (RADWAG, Radom, Poland). The concentrations of glucose, fructose, mannose, glycerol, citric acid and polyols were determined using the Dionex UltiMate 3000 HPLC instrument (Dionex-Thermo Fisher, Sunnyvale, CA, USA) equipped with a Carbohydrate H+ column (Thermo Scientific) coupled to UV (λ = 210 nm) and RI (Shodex, Ogimachi, Japan) detectors. The column was eluted with 25 mM trifluoroacetic acid (TFA) at 65 ◦C and a flow rate of 0.6 mL/min. A two-step direct transesterification method [47] was used for lipid quantification. Briefly, 1 mg of biomass was collected, centrifuged for 2 min at 8000 rpm and the super- natant was discarded. Next, 100 µL of C17:0 internal standard (Sigma-Aldrich, St. Louis, MO, USA) was added to the samples. After the addition of 500 µL 0.5 M sodium methoxide, the samples were vortexed for 60 min at room temperature. Following incubation, 40 µL anhydrous H2SO4 and 1 mL hexane were added and left overnight at room temperature. After centrifugation at 8000 rpm for 1 min, the upper hexane layer containing fatty acid methyl esters (FAMEs) was collected and analyzed by GC-MS instrument (Shimadzu, Kyoto, Japan) equipped with a Zebron ZB-FAME capillary column (30 m × 0.25 mm × 0.20 µm). The samples (1 µL at 250 ◦C) were injected in splitless mode using helium (1 mL/min). Identification of fatty acids was carried out by a comparison of retention times with reference compounds (Supelco 37 Component FAME Mix, Sigma-Aldrich).

4.7. Gene Expression Analysis RNA from the collected samples was immediately extracted using the Total RNA Mini kit (A&A Biotechnology, Gdynia, Poland) according to the supplied protocol and its concentration and quality was verified using the Biochrom WPA Biowave DNA spectrophotometer (Biochrom Ltd., Cambridge, UK) Extracted RNA samples were treated with DNase (A&A Biotechnology, Gdynia, Poland) and reverse transcribed to cDNA using Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Scientific). The obtained cDNA samples were then used for qPCR reaction using the Maxima SYBR Green qPCR Master Mix (Thermo Scientific) and primers listed in Table S1 in the PCRmax Eco 48 thermal cycler (Illumina, San Diego, CA, USA). The expression of analyzed genes was standardized to the expression of the actin (YlACT1, YALI0D08272g) gene. The gene expression levels were examined in three biological replicates. Int. J. Mol. Sci. 2021, 22, 9282 17 of 20

4.8. Calculations Substrate utilization rates (r), expressed as gram of substrate per liter per hour, were calculated for linear utilization curves using following equation:

S − S r = f i (1) tf − ti

where Si and Sf are substrate concentrations at the beginning and end of the linear substrate utilization, respectively, [g/L] and tf and ti represent time of culture at the beginning and end of the linear substrate utilization, respectively [h]. Maximum growth rates (gram of biomass per liter per hour) were calculated as follows: Xf − Xi rxmax = (2) tf − ti

where Xi and Xf are biomass concentrations at the beginning and end of the logarithmic growth phase, respectively, [g/L] and tf and ti represent time of culture at the beginning and end of the logarithmic growth phase, respectively [h]. The biomass yield from a gram of substrate (YX/S) was determined using:

X Y = (3) S

The mass yield coefﬁcient of product formation (YP/S), expressed as grams of product per gram of consumed substrate, was calculated from the following:

P Y = (4) S The speciﬁc productivity (q, grams of product per gram of biomass per hour) was calculated with: P q = (5) X ∗ t The volumetric productivity (Q, grams of product per liter per hour) was calculated using: P Q = (6) V ∗ t where P is the total amount of product in the culture medium at the end of cultivation [g], S is the total amount of the consumed substrate [g], V is the initial volume of culture [L], X is the biomass concentration [g] and t is the time of cultivation [h].

5. Conclusions In the present study, we intended to point out molecular actors responsible for preferential glucose over fructose uptake in the oleaginous yeast Y. lipolytica. Specifically, we chose proteins responsible for sugar transport and activation, and by combining a variety of experimental methods, including heterologous expression, time-course experiments and bioreactor studies, we showed that the interplay between hexose transporters and hexokinase is implicated in the preferred glucose consumption. However, in order to recognize the true nature of these interactions on glucose preference, future studies with Y. lipolytica mutants devoid of glucokinase are necessary. Additionally, we indicated that the input flux of carbon has a large impact on downstream metabolic pathways and that this flux can be directed to the molecules of industrial significance.

Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ijms22179282/s1. Author Contributions: Conceptualization, A.-M.C.-L.C., J.-M.N. and Z.L.; methodology, P.H., Z.L.; validation, P.H. and Z.L.; investigation, P.H., P.S., T.W. and Z.L.; resources, P.H. and Z.L.; data Int. J. Mol. Sci. 2021, 22, 9282 18 of 20

curation, P.H. and Z.L.; writing—original draft preparation, P.H.; writing—review and editing, P.H., A.-M.C.-L.C., J.-M.N. and Z.L.; visualization, P.H. and P.S.; supervision, A.-M.C.-L.C., J.-M.N., Z.L.; project administration, Z.L.; funding acquisition, P.H. and Z.L. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Wrocław University of Environmental and Life Sciences (“Innowacyjny Doktorat”, grant no. D220/0006/17) and the APC was funded by the Leading Research Groups support project from the subsidy increased for the period 2020–2025 in the amount of 2% of the subsidy referred to Art. 387 (3) of the Law of 20 July 2018 on Higher Education and Science, obtained in 2019. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Acknowledgments: P.S. received the financial support in the form of “BioTechNan—Interdisciplinary Environmental Doctoral Program KNOW in the field of Biotechnology and Nanotechnology”. The project is co-financed by the European Union—European Social Fund—3.2 Doctoral studies of the Operational Program Knowledge, Education and Development 2014–2020. We would like to thank Eckhard Boles (Goethe-Universität Frankfurt a.M.) for S. cerevisiae strain EBY.VW4000, and Markus Künzler (ETH Zurich) for the plasmids we used to express genes in S. cerevisiae. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Ledesma-Amaro, R.; Nicaud, J.M. Yarrowia lipolytica as a biotechnological chassis to produce usual and unusual fatty acids. Prog. Lipid Res. 2016, 61, 40–50. [CrossRef][PubMed] 2. Szczepańska,P.; Hapeta, P.; Lazar, Z. Advances in production of high-value lipids by oleaginous yeasts. Crit. Rev. Biotechnol. 2021, 1–22. [CrossRef][PubMed] 3. Blazeck, J.; Hill, A.; Liu, L.; Knight, R.; Miller, J.; Pan, A.; Otoupal, P.; Alper, H.S. Harnessing Yarrowia lipolytica lipogenesis to create a platform for lipid and biofuel production. Nat. Commun. 2014, 5, 3131. [CrossRef][PubMed] 4. Dulermo, R.; Brunel, F.; Dulermo, T.; Ledesma-Amaro, R.; Vion, J.; Trassaert, M.; Thomas, S.; Nicaud, J.M.; Leplat, C. Using a vector pool containing variable-strength promoters to optimize protein production in Yarrowia lipolytica. Microb. Cell Fact. 2017, 16, 31. [CrossRef] 5. Gasmi, N.; Ayed, A.; Nicaud, J.M.; Kallel, H. Design of an efficient medium for heterologous protein production in Yarrowia lipolytica: Case of human interferon alpha 2b. Microb. Cell Fact. 2011, 10, 38. [CrossRef][PubMed] 6. Rymowicz, W.; Fatykhova, A.R.; Kamzolova, S.V.; Rywińska,A.; Morgunov, I.G. Citric acid production from glycerol-containing waste of biodiesel industry by Yarrowia lipolytica in batch, repeated batch, and cell recycle regimes. Appl. Microbiol. Biotechnol. 2010, 87, 971–979. [CrossRef] 7. Rakicka, M.; Biegalska, A.; Rymowicz, W.; Dobrowolski, A.; Mirończuk,A.M. Polyol production from waste materials by genetically modified Yarrowia lipolytica. Bioresour. Technol. 2017, 243, 393–399. [CrossRef] 8. Tomaszewska, L.; Rywińska,A.; Gładkowski, W. Production of erythritol and mannitol by Yarrowia lipolytica yeast in media containing glycerol. J. Ind. Microbiol. Biotechnol. 2012, 39, 1333–1343. [CrossRef] 9. Hapeta, P.; Kerkhoven, E.J.; Lazar, Z. Nitrogen as the major factor influencing gene expression in Yarrowia lipolytica. Biotechnol. Rep. 2020, 27, e00521. [CrossRef] 10. Lubuta, P.; Workman, M.; Kerkhoven, E.J.; Workman, C.T. Investigating the influence of glycerol on the utilization of glucose in Yarrowia lipolytica using RNA-Seq-based transcriptomics. G3-Genes Genom Genet. 2019, 9, 4059–4071. [CrossRef] 11. Rywińska,A.; Rymowicz, W.; Zarowska,˙ B.; Skrzypiński,A. Comparison of citric acid production from glycerol and glucose by different strains of Yarrowia lipolytica. World J. Microbiol. Biotechnol. 2010, 26, 1217–1224. [CrossRef][PubMed] 12. Yang, L.B.; Zhan, X.B.; Zheng, Z.Y.; Wu, J.R.; Gao, M.J.; Lin, C.C. A novel osmotic pressure control fed-batch fermentation strategy for improvement of erythritol production by Yarrowia lipolytica from glycerol. Bioresour. Technol. 2014, 151, 120–127. [CrossRef] [PubMed] 13. Barth, G.; Gaillardin, C. Yarrowia lipolytica. In Nonconventional Yeasts in Biotechnology; Springer: Berlin/Heidelberg, Germany, 1996; pp. 313–388. 14. Lazar, Z.; Gamboa-Meléndez, H.; Crutz-Le Coq, A.M.; Neuvéglise, C.; Nicaud, J.M. Awakening the endogenous Leloir pathway for efficient galactose utilization by Yarrowia lipolytica. Biotechnol. Biofuels 2015, 8, 185. [CrossRef] 15. Ledesma-Amaro, R.; Lazar, Z.; Rakicka, M.; Guo, Z.; Fouchard, F.; Crutz-Le Coq, A.M.; Nicaud, J.M. Metabolic engineering of Yarrowia lipolytica to produce chemicals and fuels from xylose. Metab. Eng. 2016, 38, 115–124. [CrossRef][PubMed] 16. Mano, J.; Liu, N.; Hammond, J.H.; Currie, D.H.; Stephanopoulos, G. Engineering Yarrowia lipolytica for the utilization of acid whey. Metab. Eng. 2020, 57, 43–50. [CrossRef] Int. J. Mol. Sci. 2021, 22, 9282 19 of 20

17. Lazar, Z.; Walczak, E.; Robak, M. Simultaneous production of citric acid and invertase by Yarrowia lipolytica SUC+ transformants. Bioresour. Technol. 2011, 102, 6982–6989. [CrossRef] 18. Berthels, N.J.; Otero, R.R.C.; Bauer, F.F.; Pretorius, I.S.; Thevelein, J.M. Correlation between glucose/fructose discrepancy and hexokinase kinetic properties in different Saccharomyces cerevisiae wine yeast strains. Appl. Microbiol. Biotechnol. 2008, 77, 1083–1091. [CrossRef] 19. Förster, A.; Aurich, A.; Mauersberger, S.; Barth, G. Citric acid production from sucrose using a recombinant strain of the yeast Yarrowia lipolytica. Appl. Microbiol. Biotechnol. 2007, 75, 1409–1417. [CrossRef][PubMed] 20. Lazar, Z.; Neuvéglise, C.; Rossignol, T.; Devillers, H.; Morin, N.; Robak, M.; Nicaud, J.M.; Crutz-Le Coq, A.M. Characterization of hexose transporters in Yarrowia lipolytica reveals new groups of sugar porters involved in yeast growth. Fungal Genet. Biol. 2017, 100, 1–12. [CrossRef] 21. Petit, T.; Gancedo, C. Molecular cloning and characterization of the gene HXK1 encoding the hexokinase from Yarrowia lipolytica. Yeast 1999, 15, 1573–1584. [CrossRef] 22. Hapeta, P.; Szczepańska,P.; Neuvéglise, C.; Lazar, Z. A 37-amino acid loop in the Yarrowia lipolytica hexokinase impacts its activity and affinity and modulates gene expression. Sci. Rep. 2021, 11, 6412. [CrossRef] 23. Qiang, S.; Wang, J.; Xiong, X.C.; Qu, Y.L.; Liu, L.; Hu, C.Y.; Meng, Y.H. Promoting the synthesis of precursor substances by overexpressing Hexokinase (Hxk) and Hydroxymethylglutaryl-CoA Synthase (Erg13) to elevate β-carotene production in engineered Yarrowia lipolytica. Front. Microbiol. 2020, 11, 1346. [CrossRef] 24. Lazar, Z.; Dulermo, T.; Neuvéglise, C.; Crutz-Le Coq, A.M.; Nicaud, J.M. Hexokinase—A limiting factor in lipid production from fructose in Yarrowia lipolytica. Metab. Eng. 2014, 26, 89–99. [CrossRef] 25. Petit, T.; Blázquez, M.A.; Gancedo, C. Schizosaccharomyces pombe possesses an unusual and a conventional hexokinase: Biochemical and molecular characterization of both hexokinases. FEBS Lett. 1996, 378, 185–189. [CrossRef] 26. Wieczorke, R.; Krampe, S.; Weierstall, T.; Freidel, K.; Hollenberg, C.P.; Boles, E. Concurrent knock-out of at least 20 transporter genes is required to block uptake of hexoses in Saccharomyces cerevisiae. FEBS Lett. 1999, 464, 123–128. [CrossRef] 27. Hayes, M.A.; Davies, C.; Dry, I.B. Isolation, functional characterization, and expression analysis of grapevine (Vitis vinifera L.) hexose transporters: Differential roles in sink and source tissues. J. Exp. Bot. 2007, 58, 1985–1997. [CrossRef][PubMed] 28. Schüßler, A.; Martin, H.; Cohen, D.; Fitz, M.; Wipf, D. Characterization of a carbohydrate transporter from symbiotic glomeromy- cotan fungi. Nature 2006, 444, 933–936. [CrossRef] 29. Papanikolaou, S.; Muniglia, L.; Chevalot, I.; Aggelis, G.; Marc, I. Yarrowia lipolytica as a potential producer of citric acid from raw glycerol. J. Appl. Microbiol. 2002, 92, 737–744. [CrossRef][PubMed] 30. Tai, M.; Stephanopoulos, G. Engineering the push and pull of lipid biosynthesis in oleaginous yeast Yarrowia lipolytica for biofuel production. Metab. Eng. 2013, 15, 1–9. [CrossRef] 31. Rakicka, M.; Kieroń,A.; Hapeta, P.; Neuvéglise, C.; Lazar, Z. Sweet and sour potential of yeast from the Yarrowia clade. Biomass Bioenerg. 2016, 92, 48–54. [CrossRef] 32. Workman, M.; Holt, P.; Thykaer, J. Comparing cellular performance of Yarrowia lipolytica during growth on glucose and glycerol in submerged cultivations. Amb Express 2013, 3, 58. [CrossRef][PubMed] 33. Dulermo, T.; Lazar, Z.; Dulermo, R.; Rakicka, M.; Haddouche, R.; Nicaud, J.M. Analysis of ATP-citrate lyase and malic enzyme mutants of Yarrowia lipolytica points out the importance of mannitol metabolism in fatty acid synthesis. BBA-Mol. Cell Biol. Lipids 2015, 1851, 1107–1117. [CrossRef][PubMed] 34. Wang, N.; Chi, P.; Zou, Y.; Xu, Y.; Xu, S.; Bilal, M.; Fickers, P.; Cheng, H. Metabolic engineering of Yarrowia lipolytica for thermoresistance and enhanced erythritol productivity. Biotechnol. Biofuels 2020, 13, 176. [CrossRef] 35. Mirończuk,A.M.; Furgała, J.; Rakicka, M.; Rymowicz, W. Enhanced production of erythritol by Yarrowia lipolytica on glycerol in repeated batch cultures. J. Ind. Microbiol. Biotechnol. 2014, 41, 57–64. [CrossRef][PubMed] 36. Quarterman, J.; Slininger, P.J.; Kurtzman, C.P.; Thompson, S.R.; Dien, B.S. A survey of yeast from the Yarrowia clade for lipid production in dilute acid pretreated lignocellulosic biomass hydrolysate. Appl. Microbiol. Biotechnol. 2017, 101, 3319–3334. [CrossRef] 37. Carly, F.; Steels, S.; Telek, S.; Vandermies, M.; Nicaud, J.M.; Fickers, P. Identification and characterization of EYD1, encoding an erythritol dehydrogenase in Yarrowia lipolytica and its application to bioconvert erythritol into erythrulose. Bioresour. Technol. 2018, 247, 963–969. [CrossRef] 38. Carly, F.; Vandermies, M.; Telek, S.; Steels, S.; Thomas, S.; Nicaud, J.M.; Fickers, P. Enhancing erythritol productivity in Yarrowia lipolytica using metabolic engineering. Metab. Eng. 2017, 42, 19–24. [CrossRef] 39. Janek, T.; Dobrowolski, A.; Biegalska, A.; Mirończuk,A.M. Characterization of erythrose reductase from Yarrowia lipolytica and its influence on erythritol synthesis. Microb. Cell Fact. 2017, 16, 118. [CrossRef][PubMed] 40. Mirończuk,A.M.; Biegalska, A.; Dobrowolski, A. Functional overexpression of genes involved in erythritol synthesis in the yeast Yarrowia lipolytica. Biotechnol. Biofuels 2017, 10, 77. [CrossRef][PubMed] 41. Carly, F.; Gamboa-Melendez, H.; Vandermies, M.; Damblon, C.; Nicaud, J.M.; Fickers, P. Identification and characterization of EYK1, a key gene for erythritol catabolism in Yarrowia lipolytica. Appl. Microbiol. 2017, 101, 6587–6596. 42. Lazar, Z.; Rossignol, T.; Verbeke, J.; Crutz-Le Coq, A.M.; Nicaud, J.M.; Robak, M. Optimized invertase expression and secretion cassette for improving Yarrowia lipolytica growth on sucrose for industrial applications. J. Ind. Microbiol. Biotechnol. 2013, 40, 1273–1283. [CrossRef][PubMed] Int. J. Mol. Sci. 2021, 22, 9282 20 of 20

43. Fickers, P.; Le Dall, M.T.; Gaillardin, C.; Thonart, P.; Nicaud, J.M. New disruption cassettes for rapid gene disruption and marker rescue in the yeast Yarrowia lipolytica. J. Microbiol. Methods 2003, 55, 727–737. [CrossRef][PubMed] 44. Le Dall, M.T.; Nicaud, J.M.; Gaillardin, C. Multiple-copy integration in the yeast Yarrowia lipolytica. Curr. Genet. 1994, 26, 38–44. [CrossRef][PubMed] 45. Trassaert, M.; Vandermies, M.; Carly, F.; Denies, O.; Thomas, S.; Fickers, P.; Nicaud, J.M. New inducible promoter for gene expression and synthetic biology in Yarrowia lipolytica. Microb. Cell Fact. 2017, 16, 141. [CrossRef] 46. Dulermo, R.; Dulermo, T.; Gamboa-Meléndez, H.; Thevenieau, F.; Nicaud, J.M. Role of Pex11p in lipid homeostasis in Yarrowia lipolytica. Eukaryot. Cell 2015, 14, 511–525. [CrossRef][PubMed] 47. Grifﬁths, M.J.; Van Hille, R.P.; Harrison, S.T.L. Selection of direct transesteriﬁcation as the preferred method for assay of fatty acid content of microalgae. Lipids 2010, 45, 1053–1060. [CrossRef]